扰动单细胞转录组测序数据的肿瘤药物响应类型鉴定方法
By constructing a two-layer single-cell similarity network and graph embedding algorithm, the problem of accurate identification of drug response types in tumors was solved, achieving accurate identification and visualization of tumor drug response types, and revealing the dynamic changes of tumor cells and the state relationship under drug action.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TSINGHUA UNIVERSITY
- Filing Date
- 2024-02-04
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies struggle to accurately identify primary drug resistance, acquired drug resistance, and sensitive drug response patterns in different cells within tumors. Furthermore, they fail to fully utilize the contextual information of paired samples, resulting in a complex interplay between drug response mechanisms and cellular state heterogeneity, making it difficult to accurately identify molecular changes caused by drug perturbation.
A two-layer single-cell similarity network was constructed, and the embedding vectors of tumor cells were obtained through a graph embedding algorithm. Cell subpopulations were aligned based on the migration probability matrix, and the abundance and state changes of cell subpopulations were quantitatively measured by relative abundance change, absolute abundance change, and state migration cost to identify the tumor drug response type.
It enables the systematic reconstruction of complex drug response processes in tumors at the cellular subpopulation level, accurately identifies tumor cell subpopulations with similar cell states and fates, and can identify acquired resistance, absolute primary resistance, relative primary resistance, or sensitivity. It provides a visualization method for dynamic changes in tumor number and state, and displays abundance changes such as cell proliferation, inhibition, and death.
Smart Images
Figure CN118016166B_ABST